Agentic coding
CursorBench v3.1
Cursor's own test of harder, real-world coding tasks inside a code editor. Higher is better.
Rankings
Higher is better172.9%
Claude Fable 5
Anthropic · Jun 9 2026
264.8%
Claude Opus 4.7
Anthropic · Apr 16 2026
364.3%
GPT-5.5
OpenAI · Apr 23 2026
463.8%
Claude Opus 4.8
Anthropic · May 28 2026
563.2%
Composer 2.5
Cursor · May 18 2026
661.2%
Claude Sonnet 5
Anthropic · Jun 30 2026
754.6%
GLM-5.2
Z.ai · Jun 16 2026
852.2%
Composer 2
Cursor · Mar 19 2026
949.8%
Gemini 3.5 Flash
Google · May 19 2026
1049%
Claude Sonnet 4.6
Anthropic · Feb 17 2026
1147.6%
Kimi K2.6
Moonshot AI · Apr 21 2026
1231.9%
Kimi K2.5
Moonshot AI · Jan 27 2026